Quantifying the Impact of Primary vs Secondary Ice Production on Arctic Clouds

Ice formation remains among the most poorly understood and hence poorly represented cloud processes in climate models. Primary ice production (PIP) has been recognized as a key process for the correct representation of the modeled cloud feedbacks; secondary ice production (SIP), however, is much less studied and its representation in climate models is largely ignored. In this study we examine the impact of PIP and SIP on Arctic clouds using the Norwegian Earth System model version 2 (NorESM2). We show that SIP is more critical at temperatures above -15oC, while PIP has larger impacts on ice number concentrations at colder conditions. Enhancing ice formation in our simulations at levels constrained with remote-sensing observations from Ny-Alesund (2016-2017) results in increased cloud liquid and cloud cover. This is in contrast to the general consensus that increasing ice formation is more likely to deplete supercooled liquid in mixed-phase clouds through the Wegener-Bergeron-Findeisen process and riming. However, constraining ice water content through changes in PIP results in biases in high-level cloud cover. In contrast, constraining ice content through changes in SIP and ice aggregation improves low-level and mid-level cloud cover, resulting in decreased surface longwave biases, compared to satellite retrievals, during the cold months.